Context sensitive recognition of abrupt changes in cutting process Petar B. Petrovic * , Zivana Jakovljevic 1 , Vladimir R. Milacic 2 Department for Production Engineering, Faculty of Mechanical Engineering, University of Belgrade, Kraljice Marije 16, 11120 Belgrade 35, Serbia article info Keywords: Cutting process monitoring Dynamic pattern recognition abstract This paper presents a new generic approach to real-time monitoring of abrupt changes in cutting process. Proposed method is based on hierarchical fuzzy clustering of patterns obtained from discrete wavelet transform (DWT) of acquired signals correlated with cutting force variation in time. Cutting process is naturally highly dynamical and normally consists of mixture of various dynamic phenomena related to the chip formation process and dynamical responses of machining system, workpiece and tool itself. These phenomena are characterized by different time duration. The class of phenomena related to abrupt changes during short time interval is of special importance since they correspond to the most dramatic changes in cutting process, such as various kinds of tool failure or workpiece damage or even breakage. Due to their short time duration, discovery and recognition of these phenomena is extremely difficult. To solve given problem we have chosen DWT, fuzzy clustering and finite state automata as a formal platform for its analysis. Beside its good time localization properties, DWT is, due to asymmetric and irregular shapes of wavelets, especially suitable for analysis of signals having sharp changes or even discontinu- ities. Given properties make DWT an efficient means for extraction of representative and reliable infor- mation contents, thus making good basis for extraction of discriminative and representative features (as DWT coefficients combinations) for classification that will follow. Robustness of specific pattern rec- ognition and learning may be achieved only by taking into consideration wider context. Therefore, in tool condition pattern recognition we have considered the entire context of changes in cutting process state space that precedes and appears after the phenomenon which should be recognized. The cutting process behavior and its evolution in time are considered rather then momentary state which is represented as a point in adopted feature hyperspace of classification machine. Efficiency and practical applicability of developed method is evaluated by extensive experiments in laboratory conditions. Ó 2009 Published by Elsevier Ltd. 1. Introduction Condition monitoring has bean recognized in recent years as one of the key technologies required for realization of intelligent machine tool. As pointed out by Moriwaki (1994) intelligent ma- chine tool should be capable to process ambiguous and even con- flicting sensory inputs, to utilize experience from past, and accumulate knowledge through continuous learning by generaliza- tion. Moreover, intelligent machine tool should be responsive, sit- uation aware, autonomous and driven by self decision making. Among all the various aspects of machine tool intelligence, moni- toring of cutting process and tool condition is characterized as the most important one (Byrne, Dornfeld, & Denkena, 2003). It has been predicted (Rehorn, Jiang, & Orban, 2005) that an accurate and reliable tool condition monitoring system could result in cut- ting speed increases of 10–50%, a reduction in downtime by allow- ing it to be scheduled in advance and overall increase of savings in between 10% and 40%. 3 Cutting – Fig. 1 is naturally highly dynamical process, which normally consists of mixture of various dynamic phenomena related to the chip formation (chipping, chip fraction, friction be- tween tool and workpiece, plastic deformation in workpiece and chip, collisions between chip and tool...), as well as to the dynam- ical responses of machining system, workpiece and cutting tool itself. The class of phenomena characterized by abrupt changes during short time interval is of special importance since these phe- nomena are related to the most dramatic changes in cutting pro- cess. Two characteristic conflict situations containing abrupt changes can be perceived: catastrophic tool failure and collision between tool and workpiece, machine or fixture. Third situation 0957-4174/$ - see front matter Ó 2009 Published by Elsevier Ltd. doi:10.1016/j.eswa.2009.11.053 * Corresponding author. Tel.: +381 11 3302 435, +381 63 295 350; fax: +381 11 3370 364. E-mail addresses: pbpetrovic@mas.bg.ac.yu (P.B. Petrovic), zjakovljevic@mas. bg.ac.yu (Z. Jakovljevic), vmilacic@mas.bg.ac.yu (V.R. Milacic). 1 Tel.: +381 11 3302 264, +381 62 295 975; fax: +381 11 3370 364. 2 Tel.: +381 11 3302 341; fax: +381 11 3370 364. 3 DWT – Discrete Wavelet Transform; TCM – Tool Condition Monitoring; WT – Wavelet Transform; AE – Acoustic Emission; ANN – Artificial Neural Networks; HMM – Hidden Markov Models; CWT – Continuous Wavelet Transform; FSD – Feature Space Deformation; FSA – Finite State Acceptor; RMS – Root Mean Square. Expert Systems with Applications 37 (2010) 3721–3729 Contents lists available at ScienceDirect Expert Systems with Applications journal homepage: www.elsevier.com/locate/eswa